Most trading courses teach someone's opinion about the market. This one teaches what we could actually prove — and it says so, out loud, when the evidence is thin.
Every number in these ten modules comes from testing the same strategy on 2.5 years of real broker data, with commission modelled, split into a training half and a holdout half that the decisions were never allowed to see. Several of the lessons are things we expected to be true and weren't. Those are the useful ones.
Why price reaches for a level, takes it, and turns. Where stops cluster and why that makes them a target rather than a defence.
The depth threshold that separates a real liquidity grab from an ordinary wick, and what it actually buys you: profit factor 1.73 at 80 points versus 2.26 at 160. Includes a published correction to our own earlier figures.
The same rule, same year, same instrument: profit factor 1.36 outside a specific window and 2.16 inside it, on a third of the trades. Session structure is not a refinement — it is most of the edge.
Profit factor went from 1.11 to 1.62 by moving one number. Why a tight stop gets hunted out of trades that would have worked, and why a wider stop is not more risk.
Win rate is the most over-rated number in trading. How a strategy that loses two trades out of three makes money, and why forcing the win rate up destroys it.
Sizing is a drawdown decision, not a profit decision. Monte Carlo on the same trades in a different order — and why the backtest's drawdown is the optimistic case.
The mistake that made a losing strategy look profitable: costing an FX pair at spread only. How to convert commission into points, and why wide stops make a strategy broker-agnostic.
Holdout testing, sample size, and the multiple-comparisons trap. A filter that improved results in seven configurations out of seven — and still failed. The single most valuable module here.
The worst losing run on record is ten, and it wins about one trade in three. What that feels like, why it is not evidence of anything being broken, and the specific decision that turns a working strategy into a losing account.
Setting up the indicator, the one timezone setting everyone gets wrong, bar-close alerts, and the minimum account size the maths actually allows.
Four modules are open to everyone, including the most valuable one (module 8). If you read those and decide you don't need the rest, that is a completely reasonable outcome — you will still be a better trader than when you arrived, which is the point.
So you can judge the source rather than take it on faith:
| What was tested | Detail |
|---|---|
| Instrument & timeframe | XAUUSD, 15-minute |
| Sample | ~2.5 years of real broker bars (59,451 candles) |
| Method | Train on the first 70%, confirm on a 30% holdout never used for any decision |
| Costs | Spread plus commission; overnight swap measured separately |
| Trades in the final sample | 123 — a small sample, and we say so throughout |
| Live track record | Being built now. Results shown are backtested and from a demo account. Hypothetical results have real limitations and we do not present them as live performance. |